Upload
others
View
3
Download
0
Embed Size (px)
Citation preview
Assessment of NUCAPS S-NPP CrIS/ATMS Sounding Products Using Reference and Conventional Radiosondes
Bomin Sun1,2, Tony Reale2, Frank Tilley1,2, Mike Pettey1,2, Nick Nalli1,2 and Chris Barnet3
1 I. M. Systems Group, Inc., Rockville, Maryland 2 NOAA/NESDIS/Center for Applications and Research (STAR), College Park, Maryland 3 Science and Technology Corp., Columbia, MD
1
NOAA/STAR JPSS Annual Science Team Meeting August 8-12, 2016 College Park, MD
Outline • The NOAA Products Validation System (NPROVS) & and its
expansion (NPROVS+)
• Discussion on uncertainty arising from – Time/space mismatch in radiosonde and satellite observations – Radiosonde measurement accuracy – Radiosonde and satellite vertical resolution differences
• Analysis of S-NPP CrIS/ATMS temperature and water vapor
retrievals (IR+MW) based on collocations with Reference ( 3-yr) and conventional RAOBs ( 6 mons). – Global – Individual sites
NPROVS
Collocation Archive
NPROVS NPROVS+
Conventional Radiosondes
+GFS
GRUAN & Dedicated
Radiosondes
NUCAPS
S-NPP
NUCAPS MetOp-A MetOp-B
MIRS S-NPP
NOAA-18,19 MetOP-A MetOp-B
DMSP F18
AIRS v.6
EUMETSAT IASI
MetOp-A MetOp-B
GOES
ATOVS NOAA18,19
MetOp-A MetOP-B
COSMIC (UCAR)
GRAS MetOP-A MetOP-B
NPROVS+ Collocation
Archive
FTP VALAR
FTP
Algorithm Development
Visualization Tools: ODS
PDISP NARCS
PROCESSING
OUTPUTS
NOAA Products Validation System: NPROVS and NPROVS+
ECMWF
3
Typical NPROVS Global Collocations (1000 per day) 4
N P R O V S
Vaisala RS92 (28%), Vaisala RS41 (6%)
5
Global Climate Observing System (GCOS) Reference Upper Air Network (GRUAN)
JPSS Funded Dedicated RAOB • DOE ARM (SGP, NSA, ENA)
SSEC/Madison … (2) per week dual vs single, etc
• AEROSE • CALWATER • El Nino Rapid Response
• GRUAN processed
Ongoing coordination with “other” field experiments particularly if synchronized with S-NPP • Sterling Test Site • ARM Mobile Sites • CIRA/CSU
N P R O V S +: Reference RAOBs in satellite cal/val
GRUAN
6 GRUAN and JPSS funded Dedicated (S-NPP) RAOB Sites Of 23,600 RAOBs, 5,600 are synchronized (1373 via JPSS/ARM) since Jan 2013 thru Jun 2016
N P R O V S +
Jan-Feb 2013 AEROSE Nov-Dec 2013 AEROSE Nov-Dec 2015 AEROSE
Feb-Mar 2016 ENRR
Jan-Feb 2015 CalWater/ACAPEX
GRUAN/Dedicated radiosonde sites and date ranges January 2013 to mid-July 2016 (01)
RAOB Site Date Range & Number of launches Active Ascension Island ARM site (ASCENS, 80)
2016-04-29 to (175)
yes
CIRA/CSU (CIRA, 114)
2016-05-06 to 2016-06-28 (18)
No
AWARE Antarctic ARM site (AWARE, 80)
2015-12-04 to 2016-01-18 (169)
No
Barrow, AK ARM site (70027, 272, 81, 80)
2013-07-01 to (2579)
Yes
Beltsville, MD (BELTSV,114, 80, 272)
2013-07-01 to (178)
Yes
Boulder, CO (BOULDE, 272)
2013-07-09 to (147)
Yes
Cabauw, Netherlands (06260, 272)
2013-07-01 to (1201)
Yes
Kritimati Island ENRR (CXENRR, 80)
2016-01-26 to 2016-03-13 (96)
No
Darwin, Australia ARM site (94120, 80)
2014-04-01 to 2015-01-14 (714)
No
Eastern North Atlantic Azores ARM site (GRACIO, 80)
2013-09-28 to (2086)
Yes
Lauder, New Zealand (LAUDER, 123)
2013-07-03 to (234)
Yes
Lindenberg, Germany (10393, 272)
2013-07-01 to (4357)
Yes
Manus Island, Papua New Guinea ARM site (92036, 272, 80)
2013-07-01 to 2014-07-06 (110)
No
McMurdo, Antarctica ARM site (89664, 80)
2015-11-30 to 2016-03-31 (364)
No
GRUAN/Dedicated radiosonde sites and date ranges January 2013 to mid-July 2016 (02)
RAOB Site Date Range & Number of launches Active Nauru Island ARM site (91532, 272)
2013-07-01 to 2013-08-26 (717)
No
Ny-Alesund, Norway (01004, 272)
2013-07-01 to (1270)
Yes
Oliktok Point, AK ARM site (OLIKTO,80)
2013-09-10 to (944)
Yes
Payerne, Switzerland (06610, 272)
2013-07-02 to (106)
Yes
Potenza, Italy (16300, 272)
2013-12-19 to (73)
Yes
La Reunion Island, France ARM site (REUNIO, 272) Indian ocean away from Africa
2015-05-05 to 2016-05-28 (19)
No
San Cristobal Island, Ecuador (84008, 272) sea terrain
2013-07-26 to 2015-01-26 (142)
No
Southern Great Plains, OK ARM site (74646, 272, 80)
2013-07-01 to (4740)
Yes
Juan Santamaria, Costa Rica (78762, 272)
2013-07-11 to 2014-02-21 (39)
No
Sodankyla, Finland (02836, 272)
2013-07-03 to (2250)
Yes
Sterling, VA 71000 (81) 72403 (182) 72000 (152)
2015-10-28 to (724)
Yes
Tateno, Japan (47646, 272)
2013-07-01 to (296)
Yes
Table Mountain Facility, CA (TMFJPL, 272)
2014-12-05 to (27)
Yes
Pacific Missile Range Facility, HI (91162, 80)
2014-04-11 to 2014-04-26 (23)
No
GRUAN/Dedicated radiosonde sites and date ranges January 2013 to mid-July 2016 (03)
RAOB Site Date Range & Number of launches Active
Eureka, Northwestern terrority of Canada (71917, 272)
2013-01-10 to 2013-02-13 (17)
No
AEROSE Jan-Feb 2013 2013-01-09 to 2013-02-13 (109)
No
AEROSE Nov-DEC 2013 2013-11-12 to 2013-12-07 (96)
No
AEROSE Jan-Feb 2015 2015-11-17 to 2015-12-13 (90)
No
CalWater/ACAPEX Jan-Feb 2015 2015-01-12 to 2015-02-10 (171)
No
ENRR Feb-Mar 2016 2016-02-16 to 2016-03-16 (166)
No
Roles of NPROVS in sounding EDR cal/val
• Routine product monitoring (e.g., anomaly/outlier detection and long-term stability).
• Characterize product performance in a variety of meteorological conditions.
• Identify problem areas in retrieval algorithm in support of algorithm development.
• Provide independent oversight for operational product implementation.
• Provide support to AWIPS for NUCAPS applications in severe weather detection and prediction.
10
Outline • The NOAA Products Validation System (NPROVS) & and its
expansion (NPROVS+)
• Discussion on uncertainty arising from – Time/space mismatch in radiosonde and satellite observations – Radiosonde measurement accuracy – Radiosonde and satellite vertical resolution differences
• Analysis of S-NPP CrIS/ATMS temperature and water vapor
retrievals (qc-accepted IR+MW) – Global collocations – Individual sites
Satellite-RAOB Time Mismatch Impact RMS error changes with time mismatch
12
• Based on the analysis of 3-yr global IASI-RAOB collocations (506,354)
T MR RH
0.12 K/hr
2.5%/hr 1.0%/hr
Satellite-RAOB Distance Mismatch Impact RMS error changes with distance mismatch
13
• Based on the analysis of 3-yr global IASI-RAOB collocations (506,354)
MR RH T
Radiosonde temperature radiation bias impact
14
NUCAPS S-NPP IR+MW - minus - RAOB
Sample: All Data (202,00) Night (101,00) Day (6,200)
0.3 K “cooler” for Day than Night
• Radiosondes tend to have a radiation induced warm bias in the UTLS during daytime.
• The RAOB T bias at 10-70 hPa is 0.18 K for all-the –day and 0.39 K for daytime (Sun et al., JGR 2013)
Radiosonde Humidity Dry Bias Impact (Vaisala RS92 as an example vs S-NPP NUCAPS IR+MW)
15
NUCAPS-minus-RAOB RH NUCAPS-minus-RAOB MR %
Sample: All Data (8,450) Night (2,100) Day (3,050)
• Most radiosondes tend to have a dry bias in the UTLS particularly daytime.
• The RAOB RH bias at 300 hPa is ~7% for Day and ~3% (Sun et al., JGR 2011)
RS41 Improvement over RS92
16
20 dual launches of RS92 & RS41 at SGP
RS41-minus-RS92 MR %
RS41 is wetter/better
S-NPP IR+MW - minus - RAOB MR (%)
bias RMS
Sample: RS92 (6641) RS41: (1810)
Radiosonde vs. Satellite Vertical Sensitivity Impact (example of Temp inversion)
17
RMS error for the low layer: 2.2 K (weak inv) 3.0 K (strong inv)
Temp. RMS error
Sample: Strong inv (13,500) Weak inv (10,500)
Datasets used for S-NPP NUCAPS retrieval analysis
• NUCAPS-RAOB collocation data – Time mismatch: <1 hr – Distance mismatch: < 50 km
• Sonde types – Vaisala RS92 and RS41 (conventional) – Vaisala RS92 (Reference)
• Conventional RAOBs (NPROVS, 6 mons)
– 14, 000 (global), 255 (sea)
• Reference RAOBs (NPROVS+, 3 yrs)
– 4, 200 (global), 167 (sea)
18
S-NPP NUCAPS IR+MW Temperature Statistics ( K )
Relative to Conventional RAOBs Relative to Reference RAOBs
bias RMS bias RMS
8881 (global), 101 (sea) 3074 (global), 167 (sea)
S-NPP NUCAPS IR+MW H20 Vapor MR Statistics ( % )
Relative to Conventional RAOBs Relative to Reference RAOBs
bias RMS bias RMS
8881 (global), 101 (sea) 3074 (global), 167 (sea)
Suomi-NPP NUCAPS IR+MW Temperature Statistics
Jan-Feb 2015 Calwater
Jan-Feb 2013 aerose Nov-Dec 2013 aerose
Nov-Dec2015 aerose
Feb-Mar 2016 ENRR
bias RMS
Suomi-NPP NUCAPS IR+MW Water Vapor Statistics
Jan-Feb 2015 Calwater
Jan-Feb 2013 aerose Nov-Dec 2013 aerose
Nov-Dec2015 aerose
Feb-Mar 2016 ENRR
bias RMS
NUCAPS IR+ MW vs. ECMWF analysis relative to JPSS funded field campaign ship RAOBs
Water Vapor Mixing Ratio % diff.
23
Bias RMS
Beltsville & Sterling RAOBs for S-NPP Evaluation
24
• 8 Sterling synchronized to S-NPP • 90 Beltsville synchronized to S-NPP • 4 Sterling / Beltsville synchronized
to S-NPP: • 10/10/15 7100 1733 Z • 3/18/16 7200 1730 Z • 4/15/16 7100 1705 Z • 4/18/16 7100 0700 Z
Launch Synchronizations Summary
sterling
beltsville bias RMS
S-NPP IR+MW Temperature
ARM RAOBs for S-NPP Evaluation
25
• 367 at SGP site • 341 at NSA site • 73 at ENA
RAOB Launches synchronized with S-NPP
bias RMS
S-NPP IR+MW Temperature
SGP
NSA
ENA
North and South Polar RAOBs for S-NPP Evaluation
26
• 30 at AWARE ARM site • 28 at McMurdo ARM site • 398 at Ny-Alesund, Norway
RAOB Launches within 1hr/50km of S-NPP
bias RMS
S-NPP IR+MW Temperature
AWARE McMurdo
Ny-Alesund
IR+MW NUCAPS S-NPP vs. AIRS Retrieval Statistics (Sea data, relative to Reference sondes)
Temperature Water Vapor Mixing Ratio
Sample: 60 collocations (+/- 1.5 hr & 50 km)
bias STD
STD
bias
Summary • NPROVS+ (anchored to Reference RAOBs) and NPROVS (anchored to
conventional RAOB) are complementary in support of JPSS atmospheric sounding EDR cal/val
• Analysis of satellite collocations with conventional (6 months) and with Reference RAOBs (3 yrs), done globally and at individual sites, indicated – NUCAPS IR+MW temperature and water vapor retrievals perform well
• Uncertainties were discussed in the context of hyperspectral sounder retrieval validation: – Time mismatch matters – Satellite vs. radiosonde vertical resolution inconsistency – Radiosonde accuracy including warm T and dry humidity at the upper levels
28
• END
29
NUCAPS IR+ MW vs. ECMWF analysis relative to NOAA field campaign ship RAOBs
Temperature
30
bias RMS
S-NPP NUCAPS IR+MW vs. MW-only Retrieval Statistics (Sea data; relative to Reference sondes)
Temperature Water Vapor Mixing Ratio (%)
bias RMS bias RMS
IR+MW (122), MW-only (185)